TABLE OF CONTENTS In a recent in-plant training course, the members of the class were assigned a final problem. Each person was to go into the plant and set up an experiment using the techniques that had been discussed in class. One engineer wanted to study the strain readings of glass cathode supports from five different machines. Each machine had four "heads" on which the glass was formed, and the engineer decided to take four samples from each head. She treated this experiment as a 5 4 factorial with four replications per cell. Complete randomization of the testing for strain readings presented no problem. Her model was
with
Her data and analysis appear in Table 7.1. In this model she assumed that both machines and heads were fixed, and she used the 10% significance level. The results indicated no significant machine or head effect on strain readings, but there was a significant interaction at the 10% level of significance.
| Machine | ||||||
|---|---|---|---|---|---|---|
| Head | A | B | C | D | E | |
| 1 | 6 | 10 | 0 | 11 | 1 | |
| 2 | 9 | 0 | 0 | 4 | ||
| 0 | 7 | 5 | 6 | 7 | ||
| 8 | 12 | 5 | 4 | 9 | ||
| 2 | 13 | 2 | 10 | 5 | 6 | |
| 3 | 1 | 11 | 10 | 7 | ||
| 9 | 1 | 6 | 8 | 0 | ||
| 8 | 10 | 7 | 3 | 3 | ||
| 3 | 1 | 4 | 8 | 1 | 3 | |
| 10 | 1 | 5 | 8 | 0 | ||
| 0 | 7 | 0 |
